The underlying hypothesis for this project is that quantitative FDG PET imaging of tumors is a far more sensitive approach for determining response to therapy than the conventional approach of using an anatomic measure of change in size. Our specific goals are to determine the optimum time for FDG PET imaging in patients with metastatic colon carcinoma following chemotherapy and the optimum data analysis strategy. All patients identified with metastatic colon cancer who are to receive 5-FU or irinotecan chemotherapy will undergo an FDG PET scan prior to the first cycle of chemotherapy. The quantitative measures of FDG uptake and the changes in FDG uptake at specific times after chemotherapy will be correlated with changes in CEA levels and in tumor size by CT following a full course of chemotherapy lasting 3-6 months. We will also evaluate the utility of FDG PET imaging for defining the response of metastatic colorectal or primary liver tumors to non-resectional forms of therapy, including trans-hepatic artery embolization (TAE), cryosurgical ablation, ethanol injection, and TAE followed by ethanol injection. These treatments usually leave a scar in the liver that cannot be distinguished from recurrent cancer by conventional imaging tests (CT, MRI, or ultrasound). Patients will have follow up FDG PET imaging at 3-4 weeks and at 6 months. At the same time as the PET study, all patients will undergo helical CT and, when appropriate, ultrasound or CT guided biopsy of the lesion. Long term follow up will be correlated with PET, biopsy, CT, and tumor markers. A significant part of this project will be to evaluate quantitative FDG PET imaging in these patients. Simple visual examination of FDG PET images is often sufficient to identify sites of tumor. However, it is not adequate for assessing change in uptake as a measure of response to therapy and frequently is inadequate for differentiating tumor from metabolically active normal tissue. For these reasons we intend to evaluate and compare 4 different approaches for quantitating FDG uptake: standardized uptake value (SUV) and graphical analysis (Patlack method), along with glucose metabolic rate and k3 (hexokinase rate constant) using compartmental analysis with the Sokoloff/Phelps model. Sites of tumor and normal tissue will be confirmed with biopsy or appropriate follow-up. The accuracy of FDG PET imaging for predictin response to therapy after one cycle of chemotherapy will be tested by correlating change in FDG parameters with response or lack of response, and length of survival following 3-6 months of therapy.